Every year, in North America, $7.37 trillion of B2B sales orders are still keyed in manually. This costly and time-intensive process creates costs, detracts from customer service, and harms cash flow. RPA alone can’t solve this problem. However, RPA and AI working together can.

 

The Simple Appeal of Robotic Process Automation (RPA)

Many enterprises seek to graduate from manual sales order processing to a more automated approach. They look to automation to reduce their order cycle times, eliminate errors, and free their customer service representatives (CSRs) from low-value data entry tasks. Pursuing this digital transformation, companies turn to robotic process automation (RPA) implementations.

In general, RPA has many effective uses. The growth of the RPA market is indicative of this trend. More than 85% of major enterprises around the world are leveraging RPA in some way. The RPA industry will grow from $250 million in 2016 to $2.9 billion in 2021, according to Forrester. With this upward trend, it’s no surprise that many businesses assume RPA will translate well to sales order processing. If so much money is being spent on RPA, and it’s a type of automation, it must be a smart choice, right?

However, things aren’t that simple. Where RPA is effective is with simple business processes involving straightforward, linear logic. As you add complexity, the disadvantages of RPA come into sharp focus. 

 

The Disadvantages of RPA

Gartner reports that 50% of RPA software implementations will fail to deliver a sustainable ROI by 2021. Why? Primarily because enterprises try to bring RPA to bear on processes for which RPA isn’t suited. As Forbes explains,

“RPA automates manual, human processes that are highly repetitive (i.e., ‘robotic’). The most common example is data entry or management in one form or another. In these scenarios, RPA dramatically accelerates throughput while eliminating errors and reducing costs… However, RPA is not a silver bullet when it comes to digital transformation. At this stage in its maturation, many available tools do not handle complexity well.”

Sales order processing is complex and business logic may be subtle but with critical importance for processing. The variables change. With complex business processes like this, the limitations of RPA come into play:

  • RPA customization is quite restricted. While RPA can manage to automate repetitive, fixed tasks with inputs or data formats that stay consistent, RPA doesn’t work well with unexpected or frequent changes in its code integration. (This is why McKinsey senior partners have seen several robotics programs badly delayed.)
  • RPA bots can’t keep up with updates and changes. New business rules are constantly applied when processing sales orders, and this involves a lot of updates and changes of often complex business rules. The platforms on which RPA bots interact often change, and the necessary flexibility is difficult to configure into the bot. The bots are unable to meet the update needs without reconfiguration. 
  • Troubleshooting can be tedious and time-consuming. With the problems that come with reconfiguring RPA bots, companies are forced to allocate staff to protect purchase order fulfillments. Teams end up servicing bots so much that they could have manually processed the task themselves, quicker.
  • RPA requires 24/7 maintenance for order management as bots continue to break and fail. This may result in even more technical debt, as companies struggle to assign staff resources outside of standard business hours.
  • RPA implementation for order management brings about more technical debt. Manual sales order processing was poor for productivity, but RPA doesn’t help. The work simply gets transferred to the implementation team. 

 

RPA and AI: A More Intelligent Automation

None of this is to say that there is no role for RPA. With a solution such as Conexiom, RPA and AI are brought together to cover both the simpler and more challenging parts of sales order processing. 

Because RPA works well for processes that are repetitive and straightforward, some portions of sales order processing are amenable to RPA. This is why the Conexiom solution does contain some RPA components. These components cover a subset of less complicated processes, where RPA bots are a good fit. 

However, where the platform needs to constantly update its rules and codes to accommodate complex order requirements and complex business logic, more advanced forms of automation can kick in. Core RPA components can be augmented by an AI platform in order to achieve truly touchless sales order automation. These AI components can learn complex business rules, learn and apply new rules and exceptions, and self-correct errors over time. 

AI and RPA working in tandem create the ideal platform for truly touchless sales order automation. RPA can automate all the rule-based tasks, and AI can teach itself and bridge the gaps where RPA falls short. RPA and AI can coexist and support each other to support a robust and reliable platform.